Google Data Studio 如何计算不同值的总和

Question

I have data set which contains work hours and invoices related to these work hours. Some work hours are not related to any invoices yet. Several work hours can be related to a single invoice. Example of data:

Date	Project	Hours	Price of hours	Invoice	Invoice total	Invoice hours	Invoice others
2021-05-06	Project 1	7.5	500	invoice_id_1	1500	1200	300
2021-05-07	Project 1	7.5	500	invoice_id_1	1500	1200	300
2021-05-08	Project 1	7.5	600	invoice_id_2	600	600	0
2021-05-09	Project 1	2	100	invoice_id_1	1500	1200	300
2021-05-10	Project 1	7.5	550	null	null	null	null

I want to create 3 scorecards that shows the total sum of invoices, total sum of invoice hours and total sum of invoice others.

However, if I simply have SUM(invoice_hours), the sum would in this case be 4200, but I would like it to be 1800 as three of entries are in the same invoice. I could fix this by using Price of hours instead, but that doesn't help with Invoice others.

So my question is this: is it possible to count the sum of a field, counting rows where another field is not distinct only once? So in this case only counting the first row and third row (ignoring rows 2 and 4 as the invoice_id_1 has already been counted).

For what it's worth, I'm using a community connector that I created that fetches the data from an API, so I can modify the data format / add fields if it's necessary.

Answer 1

There is the need to remove duplicated. This cannot be done in Data Studio. It would be an easy task for BigQuery (select distinct Invoice, Invoice_total from ...).

Maybe you can do this as well in your community connector?

Answer 2

If further filtering of the data is not possible at the moment ( ie: being unable to blend already blended data ) and there are few different invoice IDs then a workaround could be, albeit tedious:

sum_invoiceHours :

SUM( IF( invoice = "invoice_id_1", invoice_hours , null ) )
    / COUNT( IF( invoice = "invoice_id_1", invoice , null ) )
+ SUM( IF( invoice = "invoice_id_2", invoice_hours , null ) )
    / COUNT( IF( invoice = "invoice_id_2", invoice , null ) )

Which solves as:

Now, to prevent unusual behaviours I'd recommend throwing in IFNULL() in each quotient:

IFNULL( SUM( IF( invoice = "invoice_id_1", invoice_hours , null ) )
    / COUNT( IF( invoice = "invoice_id_1", invoice , null ) ) ,0) 
+ IFNULL( SUM( IF( invoice = "invoice_id_2", invoice_hours, null ) ) 
    / COUNT( IF( invoice = "invoice_id_2", invoice , null ) ) ,0)
+ IFNULL( SUM( IF( invoice = "invoice_id_3", invoice_hours , null ) ) 
    / COUNT( IF( invoice = "invoice_id_3", invoice , null ) ) ,0)

... it's not ideal (nor pretty) but it'll get the job done for now .